Genome-Wide Approaches to Detecting Stromal-Epithelial Interactions in Breast Cancer

Principal Investigator: PEPIN, FRANCOIS
Institution Receiving Award: MCGILL UNIVERSITY
Program: BCRP
Proposal Number: BC050298
Award Number: W81XWH-06-1-0311
Funding Mechanism: Predoctoral Traineeship Award
Partnering Awards:
Award Amount: $90,000.00


Breast cancer is the most frequently diagnosed cancer in North American women, with an estimated 237,000 cases diagnosed in 2004. Breast cancer is formed from epithelial cells that become malignant. A recent focus of research has been the connective tissue that surrounds those epithelial cells. In breast cancer, the connective tissue around the tumor is different from normal connective tissue. In much the same way as every tumor is different from one person to the next, the connective tissue surrounding the tumor is also unique. It has been shown that there are interactions between the epithelial and the connective component in breast cancer that help it to develop. Recent research also has shown that a mutation in fibroblasts in connective tissue can be sufficient to induce cancer formation in mice.

The transforming growth factor beta (TGF-B) is mainly known as a growth inhibitory molecule. It has several other activities including, in some cases, favoring tumor invasiveness. Changes in the TGF-B response have been found in human breast cancer. Mutations of genes involved in TGF-B signaling can create breast tumors in mouse. The specifics of its involvement in human breast cancer are largely unknown. Several potential cancer drugs are now in preclinical or in clinical trial that can block TGF-B signaling. The signaling in TGF-B, as with most signaling pathways, involves a set of proteins that interact with each other. As such, it is impossible to know the effects of TGF-B stimulation by looking at a single gene.

It is possible, after surgery to remove a breast tumor, to take a sample of the tumor and freeze it. By using a technique known as laser capture microdissection, we can extract specific types of cell from a frozen sample. From each tumor, we collect both epithelial and connective tissue cells. The surgery removes healthy tissue around the tumor to ensure that as much of the tumor as possible was removed. This allows us to collect epithelial and connective tissue cells that come from the healthy parts of the sample.

Microarray technology allows us to observe the expression levels of over 40,000 genes at once. We propose to use this data to infer existing interactions between the epithelial and connective tissue. By comparing the expression levels of genes in both the epithelial cells and the connective tissue, we can detect whether one gene is controlling another in another tissue. Preliminary results using 13 patients have demonstrated that it is possible to detect some interactions in this way. Over 100 tumors will be available for this analysis, which will allow us to detect interactions that only occur in a fraction of patients.

By collecting the known protein interactions involved on TGF-B signaling, we expect to be able to match predicted interaction to known interactions with high confidence. We can then get a clearer picture of the events related to TGF-B signaling in a given tumor. With the development of signaling pathway databases, it should be possible to extend those tools to other pathways and get a better understanding of the signaling occurring in cancer.

Those predictions can then be validated using standard laboratory techniques and compared to the clinical parameters of the tumors to see whether they correlate with known risk factors.

Our framework provides the first cell-wide assay for identifying signaling events and interactions between tissue types. Given the importance of signaling events and the microenvironment in breast cancer and other cancer types, such techniques will help to better understand the basic biology of breast cancer.

The identification of additional interactions will certainly provide a more complete catalogue of breast cancer-related genes, a better understanding of the physiological relevance of the signaling events, and a clear understanding of how different breast cancer pathways communicate within individual cell types.

The results of our research are likely to offer new insights in breast cancer biology about the different interactions occurring in different patients. This could lead to additional prognostic markers that would help differentiate aggressive tumors requiring extensive therapy from more benign cases that do not. Those insights could also lead to new therapeutic targets, which could be used on specific patients displaying a particular type of breast cancer.